Nvidia is reportedly set to launch a new version of the GeForce RTX 5090D within the next few days. Seemingly finalised as the GeForce RTX 5090D V2, the new model is actually a cut-down version of the GeForce RTX 5090 that preceded it. By contrast, the original GeForce RTX 5090D maintained the same core hardware as the global GeForce RTX 5090 (curr. $2,399 on Amazon).
To recap, recent leaks suggest that the GeForce RTX 5090D V2 will drop to 24 GB of VRAM. While this VRAM will still be GDDR7 running at 28 Gbps, Nvidia is expected to limit its memory bus to 384-bit. In other words, the GeForce RTX 5090D V2 will offer a lower memory bandwidth than the 1,792 GB/s of other RTX 5090 variants.
On top of that, the GeForce RTX 5090D V2 is believed to be shipping with the GB202-240 GPU configuration compared to the GB202-250 and GB202-300 found inside the GeForce RTX 5090D and GeForce RTX 5090, respectively. Apparently, the 21,760 CUDA cores, 2,375 5th-generation Tensor cores and a 575 W TDP from previous GeForce RTX 5090 models return, though.
In short, the GeForce RTX 5090D V2 will be an inferior version of the GeForce RTX 5090 and GeForce RTX 5090D that are already available. Nonetheless, early retail listings indicate that the new variant will launch in China later this week for between CNY 20,699 and CNY 24,999 (~$2,881/~$3,480) from add-in board (AIB) partners like Colorful. Apparently, Nvidia hopes to market the V2 at the same CNY 16,499 (~$2,296) MSRP as the original GeForce RTX 5090D too, despite the technical differences that exist between the two cards.